How can I determine if my load cell is damaged or has a mechanical problem?

What should I do if my empty weight does not return to zero? I think one of my load cells may be damaged. How should I troubleshoot to determine if this is a load cell or a mechanical problem?

A strain gauge load cell senses changes in weight by the amount of flex in the load cell.  This gauge is bonded to a point inside the load cell cavity. The gauge-bonding agent holds it in a fixed position so the strain gauge reflects the movement of the load cell. If the gauge bond is broken, the gauge does not maintain it’s relative position and will not reflect the correct load cell body position. This bond can be broken with liquid incursions, shock, and welding. Since a load cell’s purpose is to reflect movement repeatability, this is a serious problem.  Installing the best load sensors available on the market may cost more up front, but a better protected load cell will perform longer and provide service that you can count on.

The following paragraphs and the attached example will show one method to determine the cause. This example is of a vessel with three load cells at 3,000 lb capacity each. With the rating on the load cells equal to 2 MV/V, at 9000 lb and 5 volts excitation there is a calculated change of 600 lb/mV. [9000 lb/(2 Mv/V rating * 5 v excitation)]

Measure the Millivolt output of the individual load cell signals and record your findings. If you have a Hardy Controller and a IT summing junction box you can measure the load cell mv signal from the front panel of your controller. If you do not have a Hardy Controller and IT summing junction box, the signal output wires must be disconnected to read the output of that individual load cell. Leaving all the signal wire connected will only provide the summed value of the signal outputs of all connected load cells. In the following example, the readings will be:   Load cell 1 = 2.3mv, Load cell 2 = 1.8mv, Load cell 3 = 1.91mv.

Inspect the mechanical fittings and pipes installed on the tank for rubs or binds. To establish a gross zero the scale should be calibrated. The calibration used in this example is C2® Second Generation Calibration and has been verified as accurate by using test weights.

Run a series of batches, and after each completion check the weight reading and load cell signal Millivolt levels at empty. In this example, the empty vessel weight reading had shifted from zero to –585 lb. Measuring load cell 1 = 2.3mV, 2 = .91 mV, 3 = 1.91mV. The Millivolt changes to 0.91mv, about a –1 Mv shift or close to the calculated 600 lb/mV. At empty, only load cell number two changed, and the other load cell signal voltages repeated correctly.

The vessel was re-calibrated and verified to establish a new gross zero and another series of batches ran. After a few batches, a second zero shift of –100 lb was noted. Again, checking the signal Millivolt readings and load cell number two signal Millivolt level again shifted, 2 = 0.71mV, and the Millivolt readings at empty of other two load cell signal levels were unchanged.

Again, verify any flexures going to or from your scale are not restricting the vessel. If there were a flexure problem, there would have been small indications of Millivolt signal change on each load cell. One load cell may have shown a larger change, but they all would indicate some signal change in response to a mechanical change.

After performing the above test you would find it is necessary to replace load cell number two.